View PDF |
||||||||||||||||||
|
Searching for Science:The Development of Scirus
The emergence of the World Wide Web has dramatically changed the environment in which researchers and students operate. Academic and government institutes, societies, publishers and individuals, put in tremendous efforts to make their information available online. As a result, the number of information sources has exploded. This influx of information has made it increasingly difficult to find the most relevant documents when searching on the Web. Today’s Web search engines index over six billion pages, and still don’t manage to cover all parts of the Web. One might expect users to look to the library to help guide them through this overwhelming amount of information. Yet, the opposite is true. Numerous studies show that when looking for scholarly information, users increasingly turn to the simplest search tool available: Web search engines. Library portals are their second choice.
So what makes Web search engines so compelling to users?
The first most obvious reason is that Web search engines are very easy to use and extremely accessible. Users can access Web search engines anywhere, at any time.
The second and more important reason is that search engines cover parts of the Web that library tools and portals often disregard. Traditionally, library portals have focused mainly on primary literature. However, there is more information out there that users need, such as working papers, project outlines, lab results, articles posted on pre-print servers and students or researchers who’ve faced similar research issues. The World Wide Web is not so much just a place to find information but a social network through which researchers can find each other and exchange ideas. Research has become more complex and is often of a multidisciplinary nature. Where traditionally researchers could rely on the input of the few researchers in their field, they are now dependent on insights from people in related research areas, often working in disparate locations.
Elsevier recognised the importance of public Web sources early on and in 2001 launched Scirus, a science-specific search engine. Scirus offers free and easy searching for relevant and trustworthy scientific information on the Web.
What are the principles that guided Scirus’ development?
First, when creating the Scirus search engine, developers began by making the interface intuitive and user-friendly. They recognized that aesthetics can play an important role. Simple search boxes and instant results have become the de-facto standard. Anything more complicated is likely to be dismissed. This is especially true for younger students.
Second, search results mix primary literature and Web content, clearly indicating the importance of different sources. A recent JISC (The Joint Information Systems Committee) study shows students find it increasingly difficult to distinguish authoritative results from preliminary or even biased results. With many search engines, a comment on a discussion forum sometimes gets the same academic weight as an article from a respected journal. Scirus helps users distinguish journal sources from Web sources in all stages of the search process.
Third, ranking must be optimized for scientific searching. Ranking is now more important than ever. Traditionally, users would look through all results returned by a Web search engine, but today’s usage statistics of search results show that 85% of the time Scirus users do not look beyond the first search page. Scirus uses a dictionary with over 50,000 scientific terms (developed in cooperation with the Computational Linguistics Department at the University of Munich) to identify which pages on the Web are of true scientific value. These pages are ranked higher by default. Scirus uses the same scientific dictionary to assist users in refining their subsequent searches. Scirus does this by listing scientific terms occurring frequently in the first 1,000 results.
Scirus has also implemented several other methods to ensure the best possible results in science- and education-focused Web searching. Scirus focuses only on domains on the Web that are of educational and scholarly importance, excluding large parts of the Web that return irrelevant results. Scirus interprets the user query, using sophisticated linguistic analysis, distinguishing words that together form a scientific term. It then optimises the results, only returning those matching the entire scientific term.
What has been the response to Scirus?
Scirus is a widely used and highly appreciated tool. It has won several international awards including multiple Search Engine Watch awards for Best Specialty Search Engine and a nomination for the prestigious Webby Awards’ Best Science Web Site. Even more rewarding than awards and nominations is the feedback the Scirus team receives each week, some of which appears below.
- "This is a great search engine. I have only been using it for two days and am already hooked. It filters out all of the garbage and gives results that I need. The Web definitely needs a search engine like this. Thanks again." Marcus Harikian, Student, USA
- "Your site is absolutely the best scientific search for me as a professor of genetics." Doron Lancet, Professor, Israel
- "This website is unbelievable. I have been a google.com user for quite some time, at least two years, and become quite stressed from its lack of resources and definition. Scirus is what I've been missing." Tamara N. Tresvant, Student, USA
This feedback shows users appreciate a tool such as Scirus that combines old and new methods for searching and presentating results. Scirus has managed to find an optimal balance between authoritative sources such as Medline and ScienceDirect and valuable Web information, bringing to users the best of both worlds. 
How Scirus Works
